Mario Harik on XPO, Freight Operations, and Building a Trucking Giant
Mario Harik — a software engineer who was the third employee at XPO and now runs one of North America’s largest less-than-truckload carriers — explains how the engineering design process becomes a way to run a 40,000-person company, why service is a product you sell, and why ego is the ceiling that stops you learning.
Key ideas
- Engineering is a management operating system. Harik runs the company the way an engineer solves a problem: identify the goal, collect the data, define requirements, design and build a solution, then test the outcome. Strategy decomposes into a handful of big levers; each lever into roughly ten daily KPIs; each KPI into an action plan owned by a named person — with the caveat that people, unlike code, do not run on perfection.
- Service is a product, and it is sold person by person. In less-than-truckload freight the two service measures are not damaging the freight and delivering on time. XPO tracks both back to the individual who last handled a pallet, pays supervisors partly on service, and coaches rather than punishes — dock workers see live dials on their handhelds ranking their productivity and damages against peers, and AI now inspects trailer-loading photos before a door goes over the road.
- Ego is the enemy of learning. Harik’s own lesson: he thought he was the best programmer alive until MIT put him among people ‘fifty times better’. Ego, he argues, is believing you are so good that you stop learning — a ceiling that traps you at a local maximum and stops you creating more value.
- Talent is the biggest lever in the business. Hire for three things — are they good at what they do, are they serious about the work, and are they collegial (do they ‘raise the vibe of the team’). Rank people A/B/C by a test learned from Brad Jacobs: if an A-player resigns you get a pit in your stomach; if a C-player resigns you quietly feel you can upgrade.
- Capital allocation turns on the return, not the size of the cheque. When the competitor Yellow went bankrupt in 2023, XPO used market-by-market capacity data to pick 28 terminal properties and spent roughly a billion dollars — buying scarce, hard-to-permit dock capacity ahead of an industrial recovery.
Summary
The engineering mindset as a way to run a company
Harik frames engineering as a transferable discipline: the design process — identify a problem or goal, gather data, define requirements, build a solution, test it — maps directly onto running a business, where every day brings problems to solve or goals to hit. The corollary is a data-grounded operating rhythm. Strategy breaks into big levers; each lever into about ten KPIs watched daily; and for each KPI the team reads not only the first derivative (up 5%, down 2%) but the second — the slope, whether the trend is accelerating or quietly decelerating toward trouble. Every KPI carries an action plan with a named owner. The limit of the engineer’s instinct, he concedes, is people: software should be perfect and bug-free, but the human mind does not work that way, so the framework has to be married to believing in people and managing them as individuals.
Operational excellence in less-than-truckload freight
Since XPO split into three companies — spinning off the warehousing arm (GXO) and the brokerage (RXO) — the whole management team has focused on the LTL carrier, and Harik made service the strategic centre. Service is a product: a higher service level lets you sell supplemental services at higher margin and win profitable market share, because customers choose to do business with you. LTL consolidates many customers’ pallets across a network of some 300 terminals, so the two things that can go wrong are damage (from handling freight repeatedly) and lateness. XPO tracks both to the person. Compensation for dock supervisors was re-weighted toward service; damage is traced to the last handler via a photo taken on a handheld; a ‘trailer review’ — Harik’s analogy is a software code review — has the next terminal rank the previous loader’s quality; and AI is being trained on trailer-door photos to flag an unstrapped pallet before it ships. Workers see two live dials on their scanners — productivity versus peers, and damages caused — and Harik insists the tooling lands well only because it is used to coach, not to police.
People, talent, and the A/B/C test
Harik calls the team the single biggest lever for creating value. His hiring test has three axes — technical skill and intellect, seriousness and work ethic, and collegiality (kindness, humility, a thirst to learn) — probed by ‘double- and triple-clicking’ past the headline question until he can see whether a candidate truly owned an area, and cross-checked against a fifty-page dossier of every interviewer’s notes, mined specifically for the flaw that will later cause failure. He sorts incumbents A/B/C by the resignation test learned from Brad Jacobs. Feedback is redesigned to keep people learning: open with what makes them great, and ground every criticism in data rather than the subjective, so the person sees an objective gap instead of building a defensive wall. He runs meetings on pre-read material with ranked takeaways and questions, no devices, one speaker at a time, and — Bezos-style — the leader speaks last so junior voices are not anchored.
Capital allocation and the Yellow terminals
From Jacobs, Harik took two lessons above all: think big — set outsized goals because small goals yield small results — and treat capital allocation as a CEO’s core job, judged only by the return on the capital deployed, never the size of the cheque. The 2023 bankruptcy of the rival Yellow put a scarce asset on the market: large, hard-to-permit LTL terminals, some a mile-long dock. XPO mapped Yellow’s network against its own, found the markets where growth would soon outrun capacity, wrote a business plan per property, and bought 28 of them for close to a billion dollars — improving margins immediately and buying runway for an industrial recovery. An FP&A (financial planning and analysis) team instruments every initiative, listing risks and opportunities, probability-adjusting the forecast, and pairing each operator who owns a P&L with an analyst who names the levers.
Ego, learning, and clean fuel
The through-line is learning without ego. Harik is a voracious consumer of information — hundreds of pages of research each weekend, printed podcast transcripts — and spends fifteen to thirty calm minutes each morning running a decision tree through the day’s problems, training his mind to reason faster than he can speak. Complacency he fights by setting ultra-big goals; winning, he insists, is not beating rivals but realising a team’s full potential — ‘clean fuel’, growing the pie rather than dividing it. Success, for him, is anyone — himself included — reaching their full potential while remaining emotionally fulfilled.
Speakers
- Mario Harik — CEO of XPO; a software engineer by training who joined as the company’s third employee and rose to run one of North America’s largest less-than-truckload freight carriers.
- Shane Parrish — founder of Farnam Street; host of The Knowledge Project.
See also
- Shane Parrish — host
- Michael Ovitz on Power, Dealmaking, and Building CAA — another Knowledge Project builder-operator on teamwork, reading people, and the durable assets of leadership
- James Clear on Atomic Habits, Systems over Goals, and Identity-Based Change — Harik’s KPI-and-action-plan operating rhythm is the corporate analogue of Clear’s systems-over-goals